Water inrush has become one of the main engineering hazards in tunnel and underground engineering construction. A new ideal point interval recognition model for risk assessment of water inrush was proposed to accurately predict and effectively prevent the hazard. Given the complexity and uncertainty of the geological conditions of tunnel engineering, a continuous interval of a small range was used to assign the evaluation index instead of a fixed value. The positive and negative ideal points and the ideal distance measure function were improved. The fusion method of multi-index ideal distance measure interval and the risk classification standard based on ideal closeness degree was presented. The integrated weighting method combining the analytic hierarchy process (AHP) and frequency statistic method was introduced to determine the weight of the evaluation index. The AHP was improved based on the proposed 1~5 scale and triangular fuzzy theory. Considering the dynamic risk change of water inrush, a dynamic risk assessment method was established to realize the process control of the hazard including the preliminary assessment and secondary assessment. The risk-pregnant environment factors were selected to evaluate the preliminary risk before tunnel construction. In the construction of the tunnels, the environmental factors were modified and the riskcausing factors were introduced to evaluate the secondary risk. The proposed method was used to dynamically evaluate the risk of water inrush in the river-crossing section of the Yuelongmen Tunnel from Chengdu to Lanzhou Railway. The evaluation results were in good agreement with the actual situation. The method has better grade discrimination and risk identification and has a certain guiding significance for risk prevention and control of tunnel and underground engineering geological hazards.